2019 07 25 TSC Minutes

(7:00am - 8:00am PT)

via Zoom

Hyperledger is committed to creating a safe and welcoming community for all. For more information please visit our Code of Conduct: Hyperledger Code of Conduct

TSC Members

Present?

  • Arnaud Le Hors
  • Baohua Yang
  • Binh Nguyen
  • Christopher Ferris
  • Dan Middleton
  • Hart Montgomery
  • Kelly Olson
  • Mark Wagner
  • Mic Bowman
  • Nathan George
  • Silas Davis

Resources:

Announcements

Items of discussion

Project Lifecycle Task Force - Arnaud J LE HORS

  • Vote on resolutions:
    • Issue 4: Under what circumstances can a project be moved back to Incubation? to a lab?
      • Can we move them back? The vast majority of people said "no".
      • The proposal is that projects can never move back but they can stay in any stage as long as they want.
      • If a project completely fails, the path to take is to move forward to deprecation and end-of-life.
      • What happens with projects that become non-compliant with rules?
        • The TSC would take it up to get the project back into compliance.
        • We can kick them out by moving them to end-of-life, but only if the most dire of circumstances.
      • VOTE: none opposed, no abstentions, passed by unanimous voice vote.
    • Issue 5: Should all projects be required to start as a lab prior to being proposed for Incubation?
      • We introduced labs to make it easier to start projects with minimal oversight and should we force projects start as a lab?
      • The discussion was generally "no, they shouldn't have to start as a lab".
      • The proposal is that projects can start in either place.
      • VOTE: none opposed, no abstentions, passed by unanimous voice vote.
    • (Issue 7 drafting is complete but out for review to the Marketing Committee Chair)
      • How does a project get it's name? Marketing is establishing a process.
  • Discuss:
    • Issue 8: How do we deal with a new project coming in that is already a shipping product for a company?
      • We don't want to automatically give an established project a pass and bypass all of our checks.
      • The proposal is that if a project meets all of the criteria for incubation and active status, then they should be able to submit a proposal on day one to move through incubation to active status.
      • The problem is that we require some involvement in the HL community, however the exact criteria aren't written that way. They only imply that all of the infrastructure be within the context of HL instead of generic infrastructure.

CI/CD Task Force - David Huseby

  • DRAFT: CI/CD Task Force Report
  • We're not ready to make a full proposal due to ongoing investigation.
  • No two projects are alike in their needs for CI.
  • Because we haven't required all teams to use the Jenkins, many teams have built their own CI system on their own.
  • Goals trying to achieve:
    • Get the teams together on a CI system so that CA team has the ability to collect metrics and do security checks.
    • Use HL resources to reduce financial burden on the teams.
  • Short term solutions:
    • Maybe we can start reimbursing the existing CI systems.
    • Take existing Jenkins and run minions on AWS to speed it up.
    • Take CircleCI/Gitlab-CI/AZP pick one and start moving everybody over.
  • Long term solutions:
    • Kube cluster to lift and shift everybody over but that ran into technical issues.
    • Concerns with scaling up into the future and the budget is likely to increase significantly.
  • Is there any open source support for these systems?
    • Circle CI didn't offer us anything.
    • Gitlab-CI has offered some limited free seats for open source projects.
    • AZP seems there are some things that can be done to support open source projects for free.
  • We looked at all options but there are no good ones.
  • There are no obvious answers and it is a huge problem.
  • In the long term, we should probably look for getting all projects onto one CI system.

Project Pipeline Direction

  • Componentization v. Silos
  • Convergence
    • We've brought on new projects but there's been discussion about the rules around how we pick and choose new projects to accept.
    • Adding top-level DLT projects is maybe counter-productive for the effort to drive towards convergence.
    • How do we "draw a line" around each type of component and drive towards a module of pluggable components?
    • Should we have differentiating things about each product? Consensus was that we shouldn't do that because promoting competition with each other isn't helpful.
    • Maybe we still have some areas where HL isn't complete in product offerings.
    • We're still very early in figuring out enterprise DLTs. Compared to the history of databases, we're still in our infancy.

Quarterly updates

  • 2019 Q3 Identity WG
    • Q: none.
  • 2019 Q3 Fabric Report
    • Not a lot new from the pass.
    • 1.4.2 was released last week and completes the migration from kafka to raft.
    • Diversification continues to swing away from IBM to outside contributors.
    • 2.0 alpha is out and 2.0 final with a revised chaincode lifecycle and validation pipeline refactor, due out closer to the end of the year.

Upcoming items

Backlog

 Recordings